At Golden Helix, we get to see firsthand how our software enables groundbreaking research around the world. Our tools are central to real-world discoveries, from cancer studies and prenatal diagnostics to uncovering new genetic links in rare diseases. These publications from August 2025 showcase how VarSeq and VSPipeline are helping researchers push the boundaries of what’s possible—automating workflows, streamlining variant interpretation, and ensuring reliable results at scale. These aren’t just technical wins; they’re advances directly impacting clinical decisions and scientific progress.
Title: Clinical impact of single-gene vs. panel sequencing in advanced HR + /HER2− breast cancer: insights and implications
Background: HR+/HER2− breast cancer is the most common subtype, and resistance arises despite endocrine and CDK4/6 inhibitor therapies. ctDNA testing enables non-invasive detection of mutations, but assay choice directly affects clinical utility.
Objectives: To compare single-gene hotspot testing with panel sequencing for detecting PIK3CA and additional actionable mutations in advanced HR+/HER2− breast cancer.
Subjects and Methods: A prospective study analyzed 161 plasma samples from 146 patients using SiMSen-Seq (SSS), AVENIO ctDNA Expanded (77 genes), and mFAST-SeqS, with variant annotation performed in VarSeq® (Golden Helix).
Results:
- 92.6% concordance was observed between SSS and AVENIO for PIK3CA mutations.
- AVENIO identified additional alterations including ESR1 (17%), TP53 (23%), and PI3K pathway mutations (41%).
- Broader coverage expanded trial eligibility, with 19% more PIK3CA variants detected than SSS.
- Combining tumor fraction with high VAF variants improved classification of negative liquid biopsy results.
Conclusions: Panel sequencing added clinical value beyond single-gene testing by identifying additional actionable and resistance-associated mutations. Integration of tumor fraction metrics enhanced interpretation of negative results, supporting broader use of ctDNA panels in HR+/HER2− breast cancer.
“Additionally, variants with less than 10 mutated reads, variants with VAF below 0.1%, likely germline variants (VAF ∼ 50% and z-score <3), copy number variants, and sequencing artifacts were manually excluded. Filtered variants were annotated using VarSeq (Golden Helix) and classified according to their pathogenicity using OncoKB and/or Franklin (GennoX)41. ctDNA levels were assessed as VAF (percentage of sequence reads observed matching a specific DNA variant divided by the overall coverage at that locus). For each sample, we identified the mutation with the highest VAF (hVAF).”
Klocker, E.V., Dobrić, N., Graf, R. et al. Clinical impact of single-gene vs. panel sequencing in advanced HR + /HER2− breast cancer: insights and implications. npj Breast Cancer 11, 86 (2025). https://doi.org/10.1038/s41523-025-00805-z
Title: Deep genome sequencing reveals extensive genetic heterogeneity in early human placentas
Background: Placental biopsy is commonly used for prenatal genetic diagnostics, but confined placental mosaicism can cause discrepancies with the fetal genome. While comprehensive sequencing is advancing, knowledge of sequence-level concordance between placenta and fetus in early pregnancy is limited.
Objectives: To characterize spatial and temporal genetic heterogeneity in early and late placentas using deep genome sequencing and evaluate its implications for prenatal diagnostics.
Subjects and Methods: Six pregnancies across gestational ages 14–37 weeks were studied, with biopsies from four placental quadrants, fetal tissues, CVS, and/or AF. Deep genome sequencing (mean coverage 146×) and targeted sequencing validated findings. Variants were annotated and analyzed using VarSeq® (Golden Helix).
Results:
- All 24 placental biopsies contained unique postzygotic clones, comprising 14–62% of cells.
- Clonal expansions increased with gestational age, with larger clones and higher mutational burden observed in term placentas.
- Amniotic fluid matched the fetal genome, while CVS samples often reflected confined placental variants.
- Placental variants were rarely detectable in maternal plasma cfDNA, whereas fetal de novo germline variants were consistently identified.
Conclusions: Deep sequencing revealed extensive clonal heterogeneity in all early placentas, with clones expanding over gestation. While fetal de novo variants were reliably detected, placental-specific mutations highlight the need for confirmatory AF testing when mosaicism is suspected.
“Variant calls from each sample from the same pregnancy were merged into one vcf-file using BEDtools 2.31.1. The merged vcf-files were thereafter used for forced variant calling using the DRAGEN DNA somatic pipeline of all related fetal and parental samples. Variant annotation, filtering and visualization were carried out using VarSeq v2.3.0 (Golden Helix). A variant was considered postzygotic (somatic) in placenta tissue, when meeting the following filtering criteria in one of the placenta samples: Variant allele frequency (VAF) > 5%, and SQ (somatic quality) >10 in placenta sample and VAF < 3% or >20× coverage in matched fetal samples and maternal and paternal blood samples. A variant was considered fetal de novo germline when meeting the following filtering criteria: VAF > 5%, and SQ (somatic quality) >10 in fetal samples, and AF < 3% or >20× coverage in matched maternal and paternal blood samples. Only variants in regions with sequencing coverage >50× in all related placenta and fetal samples were considered. Variant calls were restricted to high-confidence regions of the genome. Specifically, low-complexity regions and areas of low mappability were excluded using stratification BED files provided by the Genome in a Bottle (GIAB) consortium (see “Data availability”).”
Miceikaite, I., Fagerberg, C., Brasch-Andersen, C. et al. Deep genome sequencing reveals extensive genetic heterogeneity in early human placentas. Nat Commun 16, 7873 (2025). https://doi.org/10.1038/s41467-025-63296-3
Title: SORL1 as a Putative Candidate Gene for a Novel Recessive Form of Complicated Hereditary Spastic Paraplegia: Insights from a Deep Functional Study
Background: Hereditary spastic paraplegia (HSP) is a heterogeneous neurodegenerative disorder often linked to defects in the endolysosome and autophagy pathway. SORL1, typically associated with dominant Alzheimer disease and cerebral amyloid angiopathy, has not been described in a recessive form of HSP.
Objectives: To investigate the pathogenicity of a biallelic splice donor variant in SORL1 in a patient with spastic paraparesis and cerebroretinal vasculopathy (CRV).
Subjects and Methods: A 22-year-old male proband underwent whole-exome and whole-genome sequencing, RNA sequencing, and functional studies. Variant annotation and CNV analysis were performed using VarSeq® (Golden Helix) following ACMG/AMP guidelines. Fibroblast immunofluorescence and Western blotting assessed cellular impact.
Results:
- WGS identified a homozygous c.1211+1G>A splice donor variant in SORL1, classified as likely pathogenic.
- RNA-seq confirmed aberrant splicing and nonsense-mediated decay, with ~2.5-fold reduction in SORL1 transcript levels.
- Western blotting showed near absence of full-length SORLA protein.
- Patient-derived fibroblasts demonstrated significantly enlarged endosomes and lysosomes, consistent with SORL1 loss of function.
Conclusions: The study provides genetic, transcriptomic, and functional evidence supporting SORL1 as a candidate gene for a recessive form of complicated HSP with CRV. Findings highlight a potential pathogenetic link between HSP and Alzheimer-related pathways.
“We used VarSeq v2.5 (Golden Helix, Inc., USA) for variant annotation, filtering, and interpretation as per the ACMG/AMP guidelines [11]. Variants were annotated against various public databases of allele frequencies and phenotype associations, sequence conservation, functional effect, and splicing effect predictions.
Copy number variations (CNVs) were called using VarSeq CNV pipeline, and CNVs with a p value of <0.0005 and a z score value greater than ±2 were considered for further evaluation. We annotated and interpreted CNVs in accordance with the ACMG guidelines. We used the VarSeq pipeline and AutoMap [12] to call copy-neutral loss of heterozygosity (LOH) regions as well as SV analysis for both the SRWGS and LRWGS datasets. Sanger sequencing was performed using the BDT v3.1 kit and SeqStudio system to validate the candidate variants.”
Kotambail, A., Markandeya, Y. S., Mahima, R., Sukrutha, R., Nimonkar, M. M., Dash, S. S., Prasad, C., Chetan, G. K., Mailankody, P., & Arunachal, G. (2025). SORL1 as a Putative Candidate Gene for a Novel Recessive Form of Complicated Hereditary Spastic Paraplegia: Insights from a Deep Functional Study. Preprints. https://doi.org/10.20944/preprints202508.0787.v1
These studies from August 2025 show how Golden Helix empowers researchers to turn complex sequencing data into actionable insights. With VarSeq® for annotation, filtering, CNV detection, and visualization, and VSPipeline for automation and reporting, teams can move confidently from raw data to meaningful conclusions. This is what drives us: advancing science, improving patient outcomes, and shaping the future of precision medicine.